1997-11-22 egcs snapshot available

1997-11-22 egcs snapshot available

[Jeffrey A Law]

egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
(and on various mirrors shortly, see the home page for mirror sites).

We expect to make one more snapshot (Wednesday) before the first release,
between now and then I'll be pouring over all the egcs testing info folks
have sent (thanks everyone!).  Hopefully we won't need to make any significant
changes between now and Wednesday.

I'll so be working on doc issues for this release most of the week.  I'm
hoping we can ship the first release by 11/30.

Changes in this snapshot:
        * Fix alpha config.guess problem, alpha-vms & alpha-nt bugs.

        * Minor fixes for systems with 14char filename limits.
        
        * Fix so g++ works with native assembler on hpux7,8,9

        * Workaround for conditional move bug on ppro.

        * More exception and flow graph fixes.

        * Fix PA codegen bug.

        * Fix FIXPROTO buglet on irix5

        * Minor CSE bugfixes.

        * Minor sco5 bugfixes.

        * Minor bugfixes for make compare/compare-lean.

        * Some MIPS codegen fixes.

        * Various C++ bugfixes.

        * Minor g77 & g77-runtime fixes.

        * Minor libio/libstdc++ bugfixes.


egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
(and on various mirrors shortly, see the home page for mirror sites).

You'll find:

  egcs-971122.tar.gz                    The full egcs snapshot, including all
                                        languages runtime libraries and
                                        testsuites.

  egcs-core-971122.tar.gz               Just the C and objc language front ends.

  egcs-g++-971122.tar.gz                The g++ language and runtime.

  egcs-g77-971122.tar.gz                The g77 language and runtime.
   
  egcs-objc-971122.tar.gz               The objc runtime.


  egcs-testsuite-971122.tar.gz          The c & g++ testusites.
  egcs-g++-testsuite-971122.tar.gz      The g++ testsuite.
  egcs-g77-testsuite-971122.tar.gz      The g77 testsuite.


Diffs to bring the 971114 to 971122 snapshot are available.

Re: 1997-11-22 egcs snapshot available

[Robert Lipe]

> egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22

Are the perms on this directory perhaps incorrect?

> I'll so be working on doc issues for this release most of the week.  I'm
> hoping we can ship the first release by 11/30.

Are you counting on having more or less help available over the holiday
weekend? :-)


RJL

Re: 1997-11-22 egcs snapshot available

[Jeffrey A Law]

  In message < 19971122222426.09096@dgii.com >you write:
  > > egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/199
  > 7-11-22
  > 
  > Are the perms on this directory perhaps incorrect?
Just a goof in the script which kept stuff from being installed correctly --
had to hack things since my machine is still off the net :(  Damn uswest.

  > > I'll so be working on doc issues for this release most of the week.  I'm
  > > hoping we can ship the first release by 11/30.
  > 
  > Are you counting on having more or less help available over the holiday
  > weekend? :-)
Less :-)  The idea is barring any major goof's Wednesday's snapshot is the
release (minus doc stuff which I'll be working on over the holiday weekend).

jeff

Re: 1997-11-22 egcs snapshot available

[Mumit Khan]

Jeffrey A Law <law@cygnus.com> writes:
> 
> egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11
> -22
> (and on various mirrors shortly, see the home page for mirror sites).

egcs-971114-971122.diff.gz misses version.c. `make install' wiped out 
my last snapshot ;-)

*** gcc/version.c.~1    Sun Nov 23 00:03:16 1997
--- gcc/version.c   	Sun Nov 23 00:03:26 1997
***************
*** 1 ****
! char *version_string = "egcs-2.90.17 971114 (gcc2-970802 experimental)";
--- 1 ----
! char *version_string = "egcs-2.90.18 971122 (gcc2-970802 experimental)";


Regards,
Mumit -- khan@xraylith.wisc.edu
http://www.xraylith.wisc.edu/~khan/

Linktime problems on linix i586

[440 Database Project]

Help,

	I have been trying to use egcs to compile DBMS/DBML
project for a class I am taking in DB design and implamentation. 
I have included the make file I wrote.  I have not upgraded to
egcs-971123 stapshot yet, I am using egcs-971114.

--> m@t ( Matthew W. Coan )
--> mindb@www.student.fsu.umd.edu

=================================================================

# M.  Matthew 
# I.  Ilya 
# N.  Nick 
# D.  Dan 
# B.  Bin
#
# mindb Or A verrey small DBMS/DBML.
#

MINDB = /my_home                  # Where the mindb code lives.
EGCS = ${MINDB}/egcs-971114       # Where the egcs compilers live.
CC = ${EGCS}/gcc/xgcc             # The experamental egcs gcc
CFLAGS = -g                       # Flags sent to the C compiler.
CXX = ${EGCS}/gcc/g++             # EGCS GNU C++ compiler.
CXXFLAGS = -g -fexceptions        # Flags sent to the c++ compiler.
CXXINCLUDES = -I${MINDB}/stl \    # Include directorys for the c++ compiler.
-I${EGCS}/libio \                 #
-I${EGCS}/libstdc++               #
CINCLUDES = -I${EGCS}/gcc/include # Include directorys for the c compiler.
                 
LIBS = -L${EGCS} \          # This is the hard part that dose not seam to 
-L${EGCS}/gcc/libgcc1.a \   # be working.  I tell the c/c++ compiler where 
-L${EGCS}/gcc/libgcc2.a \   # the libs live but it still gives something 
-L${EGCS}/gcc/libgcc.a      # like -lgcc ***File not found.
-L${EGCS}/libraries/libio/libio.a \ #######################
-L${EGCS}/libraries/libio/libiostream.a \
-L${EGCS}/libraries/libstdc++/libstdc++.a 

LEX = lex          # The lex compiler.
YACC = yacc        # The yacc compiler.
YFLAGS = -d        # Flags to send to the yacc compler.
MAKE = make        # The make program.
RM = /bin/rm -rf   # The rm program.
LD = ld            # The GNU linker.

# ======================================
# The mindb systems backend object code.
#
OBJS = CList.o Token.o lex.yy.o y.tab.o 

# ====================================
# Genarate the target ouput executable
# image "mindb". ( It dose not work yet! )
#
mindb  : ${OBJS} main.o
	${CXX} ${CINCLUDES} ${CXXFLAGS}  \
            ${CXXINCLUDES} ${LIBS} ${OBJS} main.o -o $@ 

main.o : main.cpp
	${CXX} ${CXXFLAGS} ${CINCLUDES} ${LIBS} ${CXXINCLUDES} $? -c 

y.tab.o : y.tab.c
	${CC} ${CFLAGS} ${CINCLUDES} ${LIBS} -c $?

lex.yy.o : lex.yy.c
	${CC} ${CINCLUDES} ${CFLAGS} ${LIBS} -c $?

Token.o : Token.c
	${CC} ${CINCLUDES} ${CFLAGS} ${LIBS} -c $?

CList.o : CList.c
	${CC} ${CINCLUDES} ${CFLAGS} ${LIBS} -c $?

y.tab.c : mindb.y
	${YACC} ${YFLAGS} $?

lex.yy.c : mindb.l
	${LEX} $?

clean :
	${RM} -rf *.o mindb lex.yy.c y.tab.c 

Re: Linktime problems on linix i586

[Jeffrey A Law]

  In message < Pine.LNX.3.95.971123081529.14982B-100000@www.student.fsu.umd.edu >
you write:
  > 
  > Help,
  > 
  > 	I have been trying to use egcs to compile DBMS/DBML
  > project for a class I am taking in DB design and implamentation. 
  > I have included the make file I wrote.  I have not upgraded to
  > egcs-971123 stapshot yet, I am using egcs-971114.
I _highly_ recommend you actually install the compiler isntead of
trying to use it from the object directory.

Change the -L arguments to be directory names, not actual library
names.

jeff

Re: 1997-11-22 egcs snapshot available

[Jeffrey A Law]

  In message < 199711230606.AAA06437@eh_pc11.xraylith.wisc.edu >you write:
  > egcs-971114-971122.diff.gz misses version.c. `make install' wiped out 
  > my last snapshot ;-)
My goof -- the normal snapshot procedures had to be severaly hacked because
my machine is still off the net.

I'll fix the tar.gz and diff.gz files, it'll take an hour or so to get them
rebuilt.


jeff

Re: 1997-11-22 egcs snapshot available

[H.J. Lu]

> 
> 
> egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
> (and on various mirrors shortly, see the home page for mirror sites).
> 

The permission on the directory doesn't allow anything as of
Sun Nov 23 12:05:51 PST 1997.


H.J.

Re: 1997-11-22 egcs snapshot available

[Jeffrey A Law]

  In message < m0xZibX-0004efC@ocean.lucon.org >you write:
  > The permission on the directory doesn't allow anything as of
  > Sun Nov 23 12:05:51 PST 1997.
I turned off perms on the dir while I was fixing the version.c botch.

It should be accessable now.

jeff

Re: 1997-11-22 egcs snapshot available

[H.J. Lu]

> 
>   In message < m0xZibX-0004efC@ocean.lucon.org >you write:
>   > The permission on the directory doesn't allow anything as of
>   > Sun Nov 23 12:05:51 PST 1997.
> I turned off perms on the dir while I was fixing the version.c botch.
> 
> It should be accessable now.
> 

But now it is egcs-971122-971114.diff.gz.

-- 
H.J. Lu (hjl@gnu.org)

Re: 1997-11-22 egcs snapshot available

[Robert Lipe]

> > egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
> > (and on various mirrors shortly, see the home page for mirror sites).
> > 
> 
> The permission on the directory doesn't allow anything as of
> Sun Nov 23 12:05:51 PST 1997.

You can get to the patch, but not the full thing.   And unless you 
know the full name of the patch (it's not hard to decude) you will
probably find it easier to go through the web version or just use
this URL.

ftp://ftp.cygnus.com/pub/egcs/snapshots/1997-11-22/egcs-971114-971122.diff.gz

The patch can be gotten.

RJL

Re: 1997-11-22 egcs snapshot available

[Oleg Krivosheev]

On Sat, 22 Nov 1997, Jeffrey A Law wrote:

> 
> egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
> (and on various mirrors shortly, see the home page for mirror sites).
> 

Dir 1997-11-22  has very nice permission - d---------

how we suppose to get it ?


regards

OK

Re: 1997-11-22 egcs snapshot available

[Philippe De Muyter]

> egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/1997-11-22
> [...]
> Diffs to bring the 971114 to 971122 snapshot are available.

The diff file is now named egcs-971122-971114.diff.gz
                         new version ^ ^ old version

The previous diff files were named egcs-<oldversion>-<newversion>.

Is the change intentional ?

I'd prefer to keep the old convention, because the old version name
fit in my 14-character file names, whereas with the convention used
for 971122, the old version name falls outside of the 14 char limit,
and automated scripts can't work anymore.

Re: 1997-11-22 egcs snapshot available

[Jeffrey A Law]

  In message < 199711232058.VAA16355@mail.macqel.be >you write:
  > > egcs-ss-971122 is now available in ftp.cygnus.com:/pub/egcs/snapshots/199
  > 7-11-22
  > > [...]
  > > Diffs to bring the 971114 to 971122 snapshot are available.
  > 
  > The diff file is now named egcs-971122-971114.diff.gz
  >                          new version ^ ^ old version
  > 
  > The previous diff files were named egcs-<oldversion>-<newversion>.
  > 
  > Is the change intentional ?
No it was not intentional -- another goof because I couldn't use my
normal procedures.  The dates ended up reversed.  It'll take an hour
or so for the snapshot to rebuild with the correct ordering
jeff

Re: 1997-11-22 egcs snapshot available

[Jeffrey A Law]

The reason egcs fails is because it's much better at finding loop
invariants than gcc-2.7.

As a result you end up with an asm like this:

(insn 37 34 39 (asm_operands/v ("movw $(( 1 << 3) - 1),%1
        movw %%ax,%2
        rorl $16,%%eax
        movb %%al,%3
        movb $0x82,%4
        movb $0x00,%5
        movb %%ah,%6
        rorl $16,%%eax") ("") 0[
            (symbol_ref:SI ("default_ldt"))
            (mem:QI (plus:SI (reg:SI 37)
                    (reg:SI 23)))
            (mem/s:QI (reg:SI 42))
            (mem/s:QI (reg:SI 41))
            (mem/s:QI (reg:SI 40))
            (mem/s:QI (reg:SI 39))
            (mem/s:QI (reg:SI 38))
        ]
        [
            (asm_input:SI ("a"))
            (asm_input:QI ("m"))
            (asm_input:QI ("m"))
            (asm_input:QI ("m"))
            (asm_input:QI ("m"))
            (asm_input:QI ("m"))
            (asm_input:QI ("m"))
        ]  ("k.c") 16) -1 (nil)
    (nil))

Note that this asm needs 7 hard registers, which is more than you can
count on being available on the x86.

So while we could probably disable the aggressive loop invariant detection,
it would just be papering over the real bug, which is the asm itself.

jeff